Detection of Generalized Media-Based Modulation Signals Using Multi-Layered Message Passing

Abstract

In this paper, we consider generalized media- based modulation (GMBM) where transmit antennas and RF mirrors (which are parasitic elements placed near the transmit antennas as digitally controlled scatterers) are indexed simultaneously to convey information bits. GMBM systems offer the advantages of high rates, reduced transmit RF hardware complexity, and improved performance compared to conventional multi-antenna systems. This paper focuses on the detection of large- scale GMBM signals, where maximum-likelihood detection is not feasible due to its exponential complexity in the transmit rate. Message passing based detection is a promising low-complexity approach. However, message passing over a bipartite graph is not possible in GMBM signal detection because the elements of the GMBM transmit vector are interdependent. Our new contribution in this paper is that we propose a {\em multi-layered message passing (MLMP) algorithm} that decouples the %GMBM transmit vector elements dependencies through the addition of new layers and constraints that account for the antenna activation and RF mirror activation patterns. The proposed MLMP algorithm is shown to scale well in complexity and also to achieve good bit error performance in high-rate large- dimension systems.